Mesa Board with high tolerance for high frequency with 24v Step-gen?

If you use a separate power supply. it must have floating outputs and
be closely coupled to the noise source (likely the drives ground coupled
to the table through the motor shielding/ motor frame shorted to table)
Unfortunately, now you have even more equipment
radiating EMI in your electronics cabinet...

The question is why do you have massive spikes on the drive common
relative to the 7I96S?

Are the cables from the 7I96S to the drive shielded?
Are the grounded only at the 7I96S end?

As I said before, this is a machine grounding/shielding issue.

The other issue with simple patching around the grounding issue is that you may lose
steps when you start/stop the torch

santy wrote:

I found some ADUM modules for 1$... which is fantastic. if the ever break, i can create a scenario to replace them easily.

My question here is about the 5v on the Servo side.
My servos dont have a 5v output channel. Can i use a 5v Power supply to power the servo side of the 3 ADUM modules?
 

 
I searched around a little and didn't find any 1-way ADuM modules, so I'm curious what you're looking at. Modules with an ADuM1201 have 1 input and 1 output, they are for single ended serial applications. You could use just 1 side but you'd need 2 or 3 modules for a single stepgen and that's no good.

I think trying to be "cheap" about it is a bad idea, no way to know if the chips on those little PCBs from Amazon or Ebay are even real. Fake ICs are a real problem.

Sometimes in specific applications you can get away with not powering 1 side of an ADuM chip, just use the ground. I think this really comes down to how much strength is required to drive that side. I don't think I'd recommend it but I do have one thing I do where it works so well without powering 1 side I don't even bother to. I have another where I can't get away with that.

That said you should probably post a manual for your servo drives as that drive has to have a way of powering it's own outputs.

I had some time last night so I whipped up a 6 channel and an 18 channel PCB. I have to get some other boards ran soon so I figured I'd stuff these in there.

 
 


 

PCW wrote:

If you use a separate power supply. it must have floating outputs and
be closely coupled to the noise source (likely the drives ground coupled
to the table through the motor shielding/ motor frame shorted to table)
Unfortunately, now you have even more equipment
radiating EMI in your electronics cabinet...

The question is why do you have massive spikes on the drive common
relative to the 7I96S?

Are the cables from the 7I96S to the drive shielded?
Are the grounded only at the 7I96S end?

As I said before, this is a machine grounding/shielding issue.

The other issue with simple patching around the grounding issue is that you may lose
steps when you start/stop the torch
 

I don't know if we can really compare the EMI of an extra 5v power supply (assuming he needs one and doesn't already have 5v available in that cabinet) to striking a plasma arc in open air from a high frequency start. This thread was going in the direction of sticking caps and resistors in the stepper outputs so what's wrong with a digital isolator that's already pretty similar to the one on the outputs of the 7i96 (I assume)?

If the gremlins bouncing around the shop are really that untamed he's gonna blow up stuff that's not even connected to his machine. I've never blown anything up with a handheld TIG torch or Plasma cutter but then again my TIG or plasma torch wasn't being controlled by sensitive electronics either. What are we speculating that he's actually blowing up on a 7i96 anyway.....Hex inverters? I have like 3 7i96's, they're great but I don't control a plasma cutter with them, if that card was designed specifically to control a plasma cutter would it not have isolated step outputs?

 

The capacitor and isolators are workarounds for a grounding/shielding issue.

If the EMI is so bad it damages the 7i96 driver outputs. It can damage
the servo drives and other components also. Noise of this magnitude
does not belong in the electronics cabinet at all.

Isolation does not solve the fundamental issue. Without solving the
grounding/shielding issue, you are re-arranging the deck chairs on the
Titanic...

The perfect world scenarios are always great but I think he said he did everything people here have suggested.

The assumption here I suppose has been that the plasma arc is radiating and killing his card. Probably a good assumption however VFDs and servo drives themselves are notoriously bad for radiating all sorts of shit into electrical cabinets.

Do you have any idea how many times I've seen a machine that cost a couple million dollars hit a plant floor, get wired up by an electrical contractor for somewhere around $100k.....We're talking isolation transformers on the 480 line, ground lugs in steel columns, extra ground rods in the concrete right next to the machine. Machine is supplied with Sola transformers on all it's AC lines supplying the Linear DC power supplies, shielded cables everywhere, you name it they did it......

Gohead and plug a auxilliary piece of equipment into it with a 50hp VFD in it, mind you that the manufacturers have prewired to work together.....Well soon as that cannon plug gets twisted the whole thing lights up like a Christmas tree. In this perfect world of grounding and shielding I wouldn't have had to solder a resistor to a cap and jam it in a terminal block but hey it happened, many times.

I can run around calling the engineers who designed the machine's idiots, or blame the electricians who installed the machines.......even though I can't find a damn thing they did wrong or I can just figure out what to do about it. Like why can I take a 1hp servo drive and install it on a converted mill in my basement without issue but the minute I put that same servo drive in a $1million machine installed by professionals it goes bonkers until I put a $5 Isolator IC on a UART line?

Sometimes it's just hard to answer these questions in the real world. So hey unless we wanna assume this guy who's smart enough to operate and maybe build his own plasma table can't figure out how to connect a ground wire or buy a roll of shielded Belden cable off Amazon, maybe it's not that simple.

tommylight wrote:

blazini36 wrote:

I don't know if we can really compare the EMI of an extra 5v power supply (assuming he needs one and doesn't already have 5v available in that cabinet) to striking a plasma arc in open air from a high frequency start.

What? Where is/was this?

blazini36 wrote:

This thread was going in the direction of sticking caps and resistors in the stepper outputs .....

Caps yes, resistors was the OP idea, not ours.
Capacitors main use is suppressing spikes, filtering, etc.
 

If you use a separate power supply. it must have floating outputs and
be closely coupled to the noise source (likely the drives ground coupled
to the table through the motor shielding/ motor frame shorted to table)
Unfortunately, now you have even more equipment
radiating EMI in your electronics cabinet...

Tommy, you seem like you're replying to things you haven't read. He's referring to the 5v supply the OP thought he may have to use for the isolator (but probably not).

I understand what the point of caps was.....completely. My problem here is that everybody was OK with suggesting he do that. Soon as I mentioned using a digital isolator "oh well you're using band-aids, go fix your problem first"

A cap is going to snub off some stray transients/noise, it doesn't stop it from entering the cable in the first place. Pretty much every single ethernet jack you have ever plugged a cable into has a magnetic isolator in it or behind it, they work.

I'm sitting in a chair right now, it's easy for me to say "hey guy go fix your grounds dummy", he says "OK I fixed my grounds" Now I either say "no you didn't" or I try to be helpful otherwise. All I'm saying is caps and resistors and .....gummy bears or whatever are not a better <insert descriptive here> than some form of isolation.

Here's the thing: For driver chips to be damaged with a plasma starting transient
(in the order of a few usec at most) there must be in the range of a 20V transient
between the 7I96 ground and the  servo drive ground in the same metal cabinet
This_IS_ a grounding or shielding issue. 

Is the 7I96S frame ground connected to the case?

Are the Servo drives frame grounds connected to the case?

Is the 7I96S's 5V Ground connected to the case?